Several times a day, I receive warnings messages from one of my clusters that tell Child X could not process all connections in time (some connections had to be delayed).
I wonder how I can troubleshot what's happening, and if there is any bottleneck.
Has anyone already seen this error before ?
These warning messages are issued if the traffic manager cannot process all current network events (a connection is ready to be accepted, read from or written to) within a short period of time. This is most likely because you're running out of CPU resources, either because you're suffering a large spike of traffic or because something else on the host system has preempted the zeus.zxtm processes.
TrafficScript rules that perform blocking operations (such as a DNS lookup, http request, java.run or authentication lookup) also add to the number of network events.
You could first verify that there is sufficient resources and that the host system is not swapping or using excessive CPU. Check that there are no other processes on the system that are using excessive CPU or memory. If you are running in a virtual enviornment and you can easily add CPU and memory, consider doing that.
Then you could try and correlate these error messages with traffic levels. Perhaps you are the victim of a malicious spike of traffic, or the beneficary of lots of custom to your website? There are techniques you could try to prioritize or slow down traffic spikes - please speak with our support team.
Further to this, there's an article on Riverbed's Support KB (support contract required) that discusses this - see
In summary, a glitch in version 8.0 resulted in this warning being raised falsely (i.e., when there was no reason to be concerned about performance). This was resolved in version 8.1.
If you are observing this warning on a later build of Stingray, please see https://supportkb.riverbed.com/support/index?page=content&id=S16377